Page 1
Decreasing Traffic Congestion in VANETs Using an
Improved Hybrid Ant Colony Optimization Algorithm
Elias Khoza, Chunling Tu, and Pius A. Owolawi Dept. Computer Systems Engineering of Tshwane University of Technology, Pretoria, South Africa
Email: {eliaskhoza, tclchunling}@gmail.com; [email protected] .
Abstract—Vehicular Ad-hoc Network (VANET) is a definitive
form of mobile ad-hoc network (MANET), which delivers data
communication in a vehicular environment, using wireless
transmission. Its fundamental goal is to increase the service
quality of Intelligent Transportation Systems (ITS), such as road
safety, logistics, and environmental kindliness, as well as
information interchange. Smart cities are encountering
problematic traffic congestion, particularly in developing
countries. This paper presents an Improved Hybrid Ant Colony
Optimization (IHACO) algorithm for decreasing congestion in
smart cities. The objective of the proposed scheme is to choose
a best routing path during rush hours by providing an optimal
path. The scheme also introduces the IHACO algorithm to
improve QoS for ITS. This algorithm (IHACO) differs from
other algorithms, such as particle swarm optimization (PSO), in
terms of pheromone update processes, which makes it more
efficient. Also, the ant colony hybrid routing protocol
(ACOHRP) protocol is introduced to improve the service
quality of intelligent traffic systems (ITS). It delivers
superlative efficiency through a better origination of packet
delivery ratio, throughput, and end-to-end delay. Simulation-
based testing is performed using Matlab simulation. It was
found that traffic congestion time decreased gradually when
using IHACO, unlike with other algorithms. The computed
results demonstrated that the IHACO algorithm offers improved
performance in terms of reliability, period, distance, and
throughput, compared with different algorithms presented in
this paper.
Index Terms—Ant colony optimization, vehicular ad-hoc
networks, hybrid routing protocol, hybrid optimization
algorithm
I. INTRODUCTION
Currently, VANET is gaining a great deal of attraction
within the industry as well as among the academic
research community. This well-thought-out system has
been measured as the most distinguished for improving
performance and efficiency for future transportation.
Congestion is mainly caused by substantial volumes of
traffic on the roads, together with other social activities
taking up road space. This increases the number of
accidents on the road because drivers all wish to reach
their destination at the earliest time. This study shows
that commuters can plan their trips by escaping congested
routes should traffic congestion be identified in advance.
Manuscript received March 15, 2020; revised August 1, 2020.
doi:10.12720/jcm.15.9.676-686
Other optimization algorithms, such as PSO can be used
to select the shortest route.
Ant Colony Optimization (ACO) is a swarm
intelligence (SI) method motivated by the behavior of
ants searching for food. ACO leaves behind a chemical
element recognized as a pheromone on their path, which
is sensed by other ants for the discovery of the best path
to follow. Other ants follow the path comprising
maximum pheromones in order to reach the source of the
food. In the ACO scenario, ants interconnect employing
pheromones; and the route of the journey is built via this
method. PSO has, however, the shortcoming of the
impulsive and slow speed of convergence.
This paper presents an improved hybrid ant colony
optimization (IHACO) algorithm for decreasing
congestion by lowering the overall travel time. This
algorithm is less expensive and more highly effective
than other algorithms available in VANET’s framework.
This paper also compares the ACOHRP routing protocol
with existing dynamic source routing (DSR) protocol,
based on throughput, data collision, and data dropped [1].
It also discusses the limitations, strengths, and strategies
of each category. Based on a qualitative comparison of
performance and environmental feasibility, it is shown
that the ACOHRP routing protocol is more accurate than
the DSR routing protocol [2].
The objective of this study is to build and develop a
VANET algorithm system, namely (IHACO), which can
detect traffic congestion in real-time, allowing vehicles to
choose the most reliable routing path available.
The remainder of this research paper is organized as
follows: Section II briefly describes the related work in
the VANET area. In Section III, the proposed approach is
presented in detail. In Section IV, the ACOHRP is
presented in detail. Section V presents the results of the
analysis of three different scenarios. Section VI
concludes the paper.
II. RELATED WORK
A. VANET Algorithms
An optimal solution to road traffic has become a
difficult task for researchers to realize. Such an optimal
solution would incorporate efficient vehicle movement on
roads. Algorithms such as Dijkstra handle the shortest
distance between source and destination. Also, the
Dijkstra algorithm has been accessed by many VANET
©2020 Journal of Communications 676
Journal of Communications Vol. 15, No. 9, September 2020
Page 2
researchers in order to realize an optimal traffic-
movement solution. However, there is no handy method
in the VANET system for managing prodigious traffic
and dynamic conditions. This results in traffic congestion.
VANET solutions have concomitant complications. Other
alternative methods must be applied [3]. Ant Colony
System is one of the algorithms that can be used to solve
the static routing issues in a vehicular environment. This
theory is based on artificial ants that are able to engage a
sub-optimal route for a real-time traffic problem [4]. The
scenario is centred on traffic blockage between several
connections, leading to an increase in overall travel time.
This was achieved with the aim of nodes being removed
or injected from the network, the time being directly
proportional to the distance between the nodes.
VANET is an encouraging and growing technology for
the next generation of vehicles. VANET offers a variety
of applications; however, the main concern is to discover
an efficient routing protocol that is feasible for the highly
dynamic VANET [5]. For this challenge, eight routing
protocol types have been discussed. For this paper, the
focus is on ‘Global Topology Routing Protocols.’ In the
past, swarm intelligence algorithms have been used in the
literature to solve real-time VANET routing issues. An
Ant system was proposed [6] that became a benchmark
for its other variants, including Max-Min Ant System and
Ant Colony System, for resolving static routing problems
in the VANET environment. This system was based on
the time that was directly proportional to the distance
between the vehicles, with the option to inkjet or
removed them from the network [7]. Other researchers
provided the modified version of ACO in order to adjust
the optimal path, which would take the least time and
respond positively to many problems in their respective
work [8]. PSO and MACO algorithms have previously
been utilized by numerous researchers for continuous
optimization issues. In the PSO algorithm scenario, each
particle will attempt to move to an improved position in
the solution space. Hence PSO was used for continuous
optimization problems, whereas ACO was used for
digital optimization problems [9].
B. VANET Routing Protocols [10]
1) Global Topology Routing Protocols: This kind of
routing protocol requires the topology of all vehicles, so
that information about links can be used to make routing
decisions in the VANET [10].
2) Topology Free Routing Protocols: These protocols
are based on position information for the moving nodes;
they are also known as geographic routing protocols [10].
3) Cluster-based Routing Protocols: These are
protocols based on the principle of clustering, in which
group formation and cluster-head selection determine the
process [11].
Geo-Cast Based Routing Protocols - These protocols
utilize GPS to learn about the position of nodes, and it is
a position based multicast routing [11].
4) Multicast-Based Routing Protocols: These
protocols focus on transmitting packets within specific
regions from a single source to numerous targets [12].
5) Broadcast-Based Routing Protocols: This is a
protocol with numerous benefits in VANET, such as the
distribution of traffic, emergency assistance, and weather
information, the road situation amongst vehicles, and the
supplying advertisements, messages, and unicasts for a
well-organized route [12].
6) Delay-Tolerant Routing Protocols: In the process
of avoiding congestion and complexity, this type of
routing is introduced into VANET with several partitions,
resulting in greater flexibility [13].
7) UAV-Assisted Routing Protocols: This is an
improved type of routing protocol completing the
connected sections while advancing routing, in order to
have a world-wide vision for UAVs [14].
C. The Structure and Behaviour of VANET
VANET and MANET have almost the same structure,
differing only in high mobility of nodes, making regular
technological variations [11]. Thus, a vehicle can rapidly
join or leave a group of vehicles in a short space of time,
resulting in having little connectivity. Moreover, VANET
supplies broadband connectivity and technical resolutions
with great accuracy [12]. When any vehicle enters the
cluster zone, its default status will be cluster member
(CM); and the HELLO message will be exchanged with
the cluster header (CH), as shown in Fig. 1.
1) Cluster creation
Vehicles change topology regularly, restricting the
lifespan of connections between vehicles, because of high
mobility and flexibility. [13]. This allows the vehicles to
move, based on predefined methods by road
infrastructure, as well as by traffic laws [14]. In order for
better communication of the application of specific data,
a cluster-header is nominated to construct and keep the
structure of the clustering mechanisms [15]. The road
segment has multi-lanes in which vehicles can travel in
different directions. Fig. 1. shows a single clustered
network in which node S acts as a cluster head, which
maintains the surrounding neighbours, namely, A, B, C,
D, E, F, H, I.
Vehicles that are in the same cluster connection are
able to exchange information efficiently for a period of
time at normal speed. [16]. The nodes cluster will be built
based on two standards identified in Fig. 1. If the cluster
is not yet recognized, and there are at least two vehicles
of which their speed and path are checked from Fig. 2
[17]. The selection of cluster header (CH) will depend on
the path of the route segment amongst vehicles. The path
should be suitably long to form a connection, allowing
for the exchange of information [18]. For the other cluster
standards, if the cluster has already been created, the
request will be thereafter be broadcast for integration
with path and speed, as shown in Fig. 2.
©2020 Journal of Communications 677
Journal of Communications Vol. 15, No. 9, September 2020
Page 3
Fig. 1. Connection of vehicles in clusters [19]
Fig. 2. Establish a connection for vehicle [3]
2) Cluster change for a simple node
The migration of a cluster from one vehicle to another
is conducted by the route and speed changes. If the
vehicle approaches the end of a mutual segment, it will
need to search for a new cluster, using cluster-head [19].
If the vehicle changes suddenly in terms of path or speed,
it will search for a new cluster by returning a warning
message to the old cluster. [20]. The vehicle must choose
its own successor before leaving the cluster if there is a
needs to change speed or path. As a result, it must relate
to the pathways of all vehicles. [21], [22].
3) Ant colony technique
Ant Colony (ACO) can be described as a metaheuristic
protocol influenced by the scavenging conduct of ants. It
uses the pheromone (hormone), which is deposited and
identified by ants when they pass along paths. [28].
Pheromones captivate the ants, which causes more ants to
be attracted to the same path. The ACO can point out a
selection technique that raises an issue by iteratively
trying to improve a candidate solution, with respect to a
particular measure of quality. In VANETs, ants are
denoted as special packets and rules that can be
configured based on the algorithm for the packets [28],
[29].
In order to build the infrastructure of a VANET system,
some inter-roadside units are required to begin the route
at a fixed distance. The red line indicates the
communication between vehicle and vehicle; whereas the
blue line indicates the vehicle-to-roadside communication.
Finally, the green line shows inter-roadside
communication. The RSUs are equipped with a Wi-Fi
router, a storage device, and an electronic device that acts
as a communicating platform, counting each vehicle that
passes the RSU. A Central Maintenance Database (CMD)
is a database that can store and analyse data received
from both vehicles and RSU. Each RSU located in that
particular area can communicate with the CMD. Every
node/vehicle can also communicate with the CMD
through the internet. Nowadays, the motor industry is
manufacturing vehicles with VANET-compatible
equipment, such as a visual warning signal, Wi-Fi, GPRS
capability, inter alia, in order to easily cope with the
VANET system. The VANET architecture uses the
intelligent transport system (ITS) to transfer data between
various on-board units (OBU). The entire scenario is
shown in Fig. 3.
Fig. 3. VANET scenario [4]
III. PROPOSED IMPROVED HYBRID ANT COLONY
OPTIMIZATION (IHACO) ALGORITHM
The Improved Hybrid Ant Colony Optimization
(IHACO) algorithm is based on Ant Colony (ACO), on
the assumption that all nodes maintain a constant speed.
In this case, vehicles may initially choose a different
route, but after receiving information about traffic
congestion, they can easily select an alternative route [23].
©2020 Journal of Communications 678
Journal of Communications Vol. 15, No. 9, September 2020
Page 4
In order to provide a smooth and congestion-free route,
the IHACO algorithm has incorporated both the PSO and
ACO algorithms. The two algorithms, PSO and ACO, are
then compared based on the best solutions found. The
algorithm with the best solution is allocated as the global
best solution of the VANET system [24]. The ANTS
parameters are then reset to default, in order to generate
the new solutions, using the global-best-solutions
parameters of the VANET system. Hence, the pheromone
is updated for the best solution of the system in IHACO,
in order to avoid congestion under normal conditions.
The ACO and PSO algorithms are summed up in a
discussion since they form the main components of
IHACO. An ACO is defined by the probabilistic method,
which was inspired by studying the conduct of ants. The
behavior of real ants is supposed to lead to the discovery
of a food source under normal circumstances [25]. As
soon as ants find the food source, they make estimates,
transporting some food back to the nest. When an ant
returns, it leaves a trail of pheromones on that route that
tips off other ants apropos of quantity and quality of the
food. Each Ant will then be able to select a route
according to the concentration of the pheromone deposit
[26]. After some time, should ants abandon a route, the
pheromones will evaporate. In the same way, vehicles are
represented as ants in the ACO algorithm, which deposits
pheromones on the traverse route. The congestion is
indicated by pheromones collected on the roads. Ant will
be led to avoid congestion if the amount of collected
pheromones exceeds a threshold value. The pheromone
value is associated with the movement of vehicles. If the
vehicle enters a particular road intersection, the threshold
value is incremented; and when the road is not busy for
some time, the pheromone value for that road is
decremented. The pheromone value is sometimes
changed to a higher value in order to avoid road
congestion. The pheromone threshold value is set back to
a reasonable value when road conditions become normal.
A. Particle Swarm Optimization (PSO)
The PSO algorithm is a metaheuristic protocol, based
on an optimization technique developed in 1995 by Drs
Eberhart and Kennedy. The technique was inspired by the
social behavior of fish schooling or birds flocking. The
technique searches for the optimum solution by regularly
updating generations in the subsequent iterations,
initializing a random population. The particles in the PSO
algorithm have both position and velocity, allowing for
the selection of those particles with the highest suitable
value in the whole search space.
B. IHACO Algorithm
This is an improved version of the existing ACO with
pheromone strength as a measure of traffic-congestion
reduction on the road. Each road is assigned an initial
random pheromone value. The pheromone value is
updated in order to reflect changes in the traffic as soon
as vehicles shift on the road. Vehicles use a pheromone
value to select the road with the least traffic; as a result,
reducing the traveling time of their trip. The assumption
here is that actual road conditions will serve the purpose
of experimentation. In order to accomplish actual road
conditions effectively, the PSO-modified algorithm has a
pbest parameter representing the local best solution
obtained. Alternatively, it has a gbest parameter denoting
the global best solution [27].
The IHACO algorithm has a global search ability as
well as a local search capacity, in making use of Ant
concurrently. Each VANET algorithm has its best
solution. However, the main difference between ACO,
PSO, and IHACO is the pheromone update process. In
the ACO algorithm, the pheromone is being updated by
ants in order to achieve their best path. The PSO
algorithm uses the pbest and gbest parameters for
updating the pheromone value. For the IHACO algorithm,
both scenarios, ants, as well as particles, are being
considered in the pheromone update process.
1) Pheromone initialization
The formula below demonstrates the speed of the
vehicle, together with the length.
𝑉𝑥 = 𝑉 ∗ cos 𝜃 𝑉𝑦 = 𝑉 ∗ sin 𝜃
(1)
(2)
where 𝑉𝑥 𝑎𝑛𝑑 𝑉𝑦 denotes the velocity of the vehicles for
both directions, simultaneously.
𝑎𝑥 and 𝑎𝑦 refer to the attitude and latitude acceleration
of vehicles.
The information drawn from the equations (1) and (2)
is measured from vehicle sensors as an assumption.
Time denoted by t is used for the pheromone
initialization formula, as given by equation (1) below:
𝜌𝑖𝑗 (𝑡) = 𝜌𝑖𝑗 (𝑡 − 1) + 𝐶1 ∗ 𝑙𝑒𝑛𝑔𝑡ℎ(𝑒𝑑𝑔𝑒𝑖𝑗) (3)
In the above equation, 𝜌𝑖𝑗 denotes the value of the
pheromone from vehicle 𝑖 to vehicle j, which is directly
proportional to the value of the edge. Time is denoted by
t = 0, which is the measurement of the time taken by each
vehicle to travel from source to destination. 𝐶1 denotes
the constant value of the vehicles in the [0, 1] range. The
more the edge increases, the more does the length. This
also illustrates a shorter and congestion-free route.
2) Pheromone update process
The IHACO, as explained, is an improved version of
ACO. An ant updates the pheromone value on the edges
after negotiating time taken by the vehicles to travel. This
method is given by the formula below:
𝜌𝑖𝑗 (𝑡) = 𝜌𝑖𝑗 (𝑡 − 1) + 𝐶2 ∗ (1 + ∆𝑝𝑖𝑗) (4)
where ∆𝑝𝑖𝑗 = 1
𝑁 , or otherwise = 0. In this scenario, N
represents a number of nodes, 𝐶2 is a constant value
which lies between [0, 1].
©2020 Journal of Communications 679
Journal of Communications Vol. 15, No. 9, September 2020
Page 5
The travel time includes waiting time at the
intersection, waiting at a green light, and is calculated as
the sum of the actual travel time plus the waiting time at
the lights.
𝜏𝑖𝑗 = (1 − 𝜌) 𝜏𝑖𝑗 + ∆𝜏𝑖𝑗 (5)
In the above equation, 𝜏𝑖𝑗 denotes the amount of
pheromone between path 𝑖 and 𝑗. The variable 𝜌 denotes
the amount of pheromone evaporation rate, whereas the
∆𝜏𝑖𝑗 value denotes the deposited volume of pheromone
[28]. ∆𝜏𝑖𝑗 is typically given by ∆𝜏𝑖𝑗𝑘 = 1/𝐿𝑘, only if ant
𝐾 travels on paths 𝑖 𝑎𝑛𝑑 𝑗 where, 𝐿𝑘 is the cost of the 𝐾𝑡ℎ
tour typical length. Otherwise ∆𝜏𝑖𝑗𝑘 = 0.
∆𝜏𝑖𝑗 (𝑡 + 𝑛) = 𝑝𝜏𝑖𝑗 (t) + ∆𝜏𝑖𝑗 (6)
where ∆𝜏𝑖𝑗 denotes the sum of increased pheromones at
the edges for the degree of pheromone dissipation. The
number of pheromones per circle is shown by (𝑡 + 𝑛).
3) System model for IHACO optimization
Fig. 4. Traffic flow estimation
The above graph shows the estimated optimal time for
reaching the destination. Some components are included
in the scenario in order for traffic flow to be efficient.
The CSD is the central server database responsible for
optimizing the entire route, in order for vehicles to reach
their destination as quickly as possible via the IHACO
algorithm. In this scenario, the vehicle will start from the
source and look for a congestion-free route in order to
reach the destination. It will connect to the CSD every
time it reaches an intersection; the CSD will feed the
correct information to the vehicle, providing the best
route to take. This will help a vehicle to avoid routes that
are congested, indicated by the red arrows. RSUs are
located at the intersection to ease the communication
amongst the vehicles. By employing the IHACO
algorithm, the CSD will notify vehicles not to use the
route from point A to B because it is congested; rather,
take point A to E, then C, since this is less congested. The
full scenario is shown in Fig. 4.
Algorithm1: for IHACO
1) Initialize pheromone update
2) While t < maximum value
For all particles
Generate routes using PSO algorithm
Calculate pbest for generated paths (ref equation 3)
3) Continue for loop
For all ants
Generate paths using ACO algorithm
Calculate ant-best paths
Continue for loop
Calculated gbest path among all pbest paths (ref
equation)
4) If gbest <= ant-best then
ant-best = gbest
gbest = ant-best
end while
Algorithm 2: Route best for IHACO
Input = Roadmap
Output = Routebest
Routebest = Routebest
Routebest = Routedistance
Pheromone = Initialise_Pheromone ()
While (nofinishcondition)
For i= (1 to n)
Ri = FindNewRoute (Pheromone, RoadMap)
Ridistance = distance (Ri)
If Ridistance <= Routebest (distance)
Routebest (distance) = Ri distance
Routebest = Ri
End
UpdatePheromone (Pheromone, Ri, Rbest)
End
Table I illustrates the route selection parameters based
on PSO, ACO, and IHACO protocols. The IHACO
algorithm has a YES on all the parameters, unlike the
PSO and ACO, which makes it the more suitable solution
for decreasing traffic congestion in real-time.
TABLE I: ROUTE SELECTION PARAMETERS
Schemes/
Protocols Distance
Connectivity
Level Congestion
Designed
For
PSO Yes No Yes MANET
ACO Yes Yes No MANET
IHACO Yes Yes Yes VANET
IV. ACO HYBRID ROUTING PROTOCOL (ACOHRP) FOR
VANET
This routing technique enables the vehicle to
communicate various information efficiently to other
vehicles for a certain time, depending on the average
©2020 Journal of Communications 680
Journal of Communications Vol. 15, No. 9, September 2020
Page 6
speed and route of the vehicles that are in connection. Fig.
2 represents a group of vehicles that will be checked
against the route and speed in the common-route segment.
On the other hand, Fig. 4 elaborates the broadcast of
vehicles against path and speed, respectively. Clusterhead
creation is responsible for sending an acceptance signal,
awaiting an acknowledgment [28].
Fig. 5. ACOHRP route selection [29]
The number of vehicles passed through RSU is
regularly updated in the central maintenance database
(CMD) by RSU. The difference between the two RSUs is
currently calculated by the CMD. The CMD will assume
the number of non-moving vehicles as the pheromone
level, by taking the algorithm of ACO into account. The
concentration level of pheromones will be triggered to
high if a large number of vehicles remain on a route [30].
A low number of vehicles staying on a route will trigger
the pheromone to low. The CMD will check all possible
routes with less possibility of having traffic congestion,
by comparing the number of vehicles. As a result, they
will make a decision on other vehicles selecting that route.
Fig. 5 shows the application of ACOHRP.
A hybrid architecture is an architecture that combines
both V2I and V2V communication. This architecture
includes wireless networking devices which are fixed
within communication units, such as access points,
cellular towers, etc. Hybrid architecture also consists of
vehicles that communicate by exchanging information
received from infrastructure equipment or other vehicles,
through ad hoc communication. The solid blue line
indicates the communication between the vehicles, while
the orange dotted line indicates the communication
infrastructure and vehicles, respectively, as shown in Fig.
6.
Fig. 6. Hybrid architecture [27]
Fig. 7. V2V Communication [29]
The system received information about the nearby
vehicles and the other vehicle through V2V
communication and employed it for traffic operation. The
location global positioning system (GPS) measuring
system provides the vehicle location as well as heading
angle information [29] corresponding to the tolerance of
commercial differential GPS. Fig. 7 shows the entire
global system.
1) ACOHRP structure for VANET
As mentioned earlier, the method is based on traffic
information in order to enable communication between
vehicles. To achieve this, the Ant Colony is employed to
signify knowledge and vehicular traffic information. This
section describes the architecture of the proposed system,
as shown in Fig. 5. ACOHRP consists of numerous
processes, such as configuring nodes, network
initialization, source and destination allocation, data
transmission, and performance analysis [29], [30].
2) Dynamic Source Routing (DSR)
DSR is a sensitive path-finding scheme which does not
need periodic HELLO packets and a warning signal. The
DSR protocol technique is deluging the packets within
the network by using route requests. The node responds
regarding destination and conveys route traverse in its
Cluster Header. DSR is composed of two techniques that
collaborate to permit route maintenance and route
©2020 Journal of Communications 681
Journal of Communications Vol. 15, No. 9, September 2020
Page 7
discovery [30]. Route maintenance is achieved by the
propagation of the route error message (RRER). Route
discovery is accomplished whenever a source node needs
to transfer any packet to an end-point node. The source
node starts by consulting its source cache [31]. DSR is
designed especially for usage in a multi-hop ad-hoc
network of mobile nodes. This permits a network to
behave entirely as self-organized and self-configuring
without using any existing infrastructure. This protocol
uses no periodic routing messages, therefore it avoids
large routing updates and decreases network bandwidth
overhead [31].
TABLE II: ROUTING PROTOCOLS COMPARISON
Parameters
Scenario
Routing Protocols
ACOHRP DSR
Throughput
Metro High Low
Data Dropped Metro Low High
Data Collide Metro Low High
From Table II, it can be seen that ACOHRP and DSR
routing protocols are being compared for the following
parameters: throughput, data dropped, and data collide.
The simulation parameters data dropped and dropped
collide are high for DSR, and low for ACOHRP. The
resulting analysis shows that ACOHRP routing protocols
have a higher performance for road traffic safety
measures than do the DSR routing protocols. Hence, it
can be justified that ACOHRP routing protocols have
high performance for the VANET system.
3) Improved ant dynamic source routing
ANT-DSR is a reactive technique that uses proactive
route protocols over a constant validation of its stored
routes. In this scenario, when packets are transmitted,
proactive methods are used within the network; and
reactive methods between networks [32]. The
performance metric is improved when the ant-net
algorithm is applied to the DSR protocol. This technique
increases the lifespan of a listed route in the VANET
network [32].
V. SIMULATION AND RESULTS
The simulations were executed on Matlab for
experimental purposes. Vehicles were entering randomly.
Each vehicle has been allocated a beginning and finishing
position within the VANET system. The participating
vehicle is taken out of the network after reaching its
destination position. Vehicles are using a random speed
ranging from 60 m/s to 90 m/s. The main objective of the
IHACO algorithm is to decrease travel time by avoiding a
congested route, regardless of the path selected. In this
scenario, the number of ants is used as the number of
vehicles, with particles being used as travel time. Hence,
particles and ants hunt for the best solution, informing the
corresponding pheromone. In order to achieve the best
global results, the initial best solution for ants and
particles is calculated and matched against one another.
The solution will be repeated until all the corresponding
parameters are acquired with respect to the global-best
solution.
The IHACO algorithm has a more superior metric for
optimization than the ACO and PSO algorithms. The
results obtained from the IHACO algorithm are compared
with those from the ACO and PSO under a similar
environment. The results are generated and depicted
graphically as the distance versus the number of vehicles,
which is shown in Fig. 8.
Fig. 8. Number of vehicles vs. distance covered
Fig. 9. Graphical representation of time taken vs. vehicles
It was also observed that the total distance covered by
the vehicles is improved when using the IHACO
algorithm, unlike the PSO and ACO approach. The single
goal of the IHACO algorithm is to avoid congestion and
to inform other vehicles of the road conditions. This can
be seen in Fig. 9 as the graphical representation of
IHACO algorithm compared with the PSO and ACO
algorithms for the time taken by vehicles to reach their
destination. When the number of vehicles decreases, the
IHACO algorithm shows a substantial decrease in the
©2020 Journal of Communications 682
Journal of Communications Vol. 15, No. 9, September 2020
Page 8
total travel time for a journey taken by vehicles. The
results obtained by using the IHACO algorithm shows
that the total distance covered by vehicles increases by
85%, unlike with the PSO algorithm, which is at 55%.
The distance using the IHACO algorithm decreases by
35%, unlike with the ACO algorithm, which decreases
with 20%. Hence, the IHACO algorithm reduces the
travel time by avoiding congestion in path length for the
vehicles.
Reliability – there is a probability that a link between
two vehicles will exist over a specific period of time.
TheIHACO algorithm shows better reliability
achievements compared with existing schemes − PSO
and ACO. The IHACO algorithm outperforms the other
owing to the adoption of congestion avoidance by
informing other vehicles about the road conditions (see
Fig. 10).
Fig. 10. Reliability of vehicles
Fig. 11. Comparison of throughput VS Vehicles
In this subsection, the performance parameters, as well
as the efficiency of network estimation, are discussed.
These parameters are:
Throughput – Fig. 11. Indicates the throughput
values of ACOHRP and DSR. The outcome confirms
the conclusion of the traffic-information results.
Throughput increases gradually for both protocols
with an increasing number of vehicles. It is seen that
ACOHRP out-performs DSR, having the highest
throughput.
End-to-End delay – Fig. 12. This represents the
average end-to-end delay of ACOHRP and DSR.
From the above observation, it can be seen that the
average end-to-end delay gradually improves for the
ACOHRP routing protocol compared with the DSR
routing protocol, owing to the time taken by the route
discovery mechanism. Also, with the increase of the
ACOHRP, the relaying of information packets has
taken much time and will result in incremental delay-
packets relay by hops. It is observed that ACOHRP
has the highest delay.
Packet loss – this refers to the packets dropped during
the transmission. Fig. 13 indicates that the DSR
routing protocol has dropped more packets than the
ACOHRP routing protocol.
By making use of the Matlab program, using
simulations, the results below were obtained. The next
measurements have been executed per simulations for the
routing protocols that have been concisely described −
routing overhead, end-to-end delay, and throughput. The
values in the figures are obtained by a simulation average
set for each vehicle.
Fig. 12. Comparison of end to end delay vs. the number of vehicles
Fig. 13. Number of vehicles vs. packet loss
The simulations and experimentations were performed
via a MATLAB version 2017b environment. Three (3)
CBR connections were used as network traffic.
©2020 Journal of Communications 683
Journal of Communications Vol. 15, No. 9, September 2020
Page 9
The simulation was conducted on 30 nodes. The
velocity for each node is constant, which is represented
by the C value. The two routing protocols are used for
comparison, namely, ACOHRP and DSR. The data
packet size is 512 bytes, respectively, using graphical
representation analysis. It was observed that, when traffic
volume increases, the total distance increases
concomitantly, as with PSO; however, it decreases
concomitantly, unlike with ACO.
TABLE III: SIMULATION PARAMETERS
Parameters Value / Range
Mobility model MATLAB
Network traffic connections 3 CBR connections
Simulation zone area 20m X 20m
Number of nodes 30
Fading Nil
The velocity of each node Constant
Speed Up to 40 m/s
Transmission range < 100 m
Routing protocols DSR, ACOHRP
Data packet size 512 bytes
The above table summarises the simulation parameters
as well as the range/value. The mobility model used is the
Matlab program with three (3) CNR connections as
network traffic. 30 Vehicles have been used for this
experiment with the constant velocity of each vehicle.
The traveling speed is up to 40 m/s with the transmission
range of 100m. Two routing protocols has been used for
this simulation, namely: DSR and ACOHRP.
VI. CONCLUSION
The IHACO approach is an improvement on the
existing ACO algorithm, which hybridized PSO and
ACO algorithms. This IHACO algorithm is introduced in
order to solve the issue of increased congestion on the
roads, assisting commuters by selecting the best path
during peak hours. The IHACO algorithm results were
compared with the results acquired from the PSO and the
ACO, respectively, using graphical representation
analysis. It was observed that, when traffic volume
increases, the total distance increases concomitantly, as
with PSO; however, it decreases concomitantly, unlike
with ACO.
The paper also focuses on traffic information, as well
as on ACOHRP, in order to allow suitable routing of
packets from source to their final destinations. A low
latency multipath routing structure has been introduced
per an ACO method for vehicular network
communication. The various routing protocols are
required in order to facilitate the proper routing of
packets to their final destination. The ant colony is used
in vehicles to enable the analysis of information acquired
from traffic. Furthermore, the ACOHRP has excellent
flexibility for routing in various ad hoc networks. It also
includes properties such as dynamic topology, efficient
path selection, as well as evaluation of link-transmission
quality. Three factors are measured to calculate
discovered paths, namely, routing overhead, end-to-end
delay, and throughput. This approach optimizes routing
by improving road-service performance, significantly
reducing delivery time. The results show that ACOHRP
performs better than DSR routing approaches. The new
proposed ACOHRP will be a better solution in dealing
with all kinds of traffic scenarios. The new ACOHRP
out-performs the VANET metrics in a dynamic
environment. Future scope includes a hybrid protocol
which can be developed to overcome all drawbacks of the
existing protocols.
CONFLICT OF INTERESTS
The authors declare that no conflict of interest for this
paper.
AUTHOR CONTRIBUTIONS
Author A wrote the full paper, including analysis of
data and simulation. Author B contribute the idea,
proofread the paper, correcting the contents and structure.
Author C provided valuable comments, and all authors
had approved the final version of the paper.
REFERENCES
[1] K. D. Thilak and A Amuthan, “Cellular automata-based
improved ant colony-based optimization algorithm for
mitigating DDoS attacks in VANETs,” Future Generation
Computer Systems, vol. 82, pp. 304-314, 2018.
[2] S. Saha, U. Roy, and D. D. Sinha, “Performance
comparison of various ad-hoc routing protocols of VANET
in Indian city scenario,” American International Journal of
Research in Science, Technology, Engineering and
Mathematics, vol. 5, no. 1, pp. 49–54, February 2014.
[3] P. Mutalik, S. Nagaraj, J. Vedavyas, R. V. Biradar, and V.
G. Patil, “A comparative study on AODV, DSR and DSDV
routing protocols for Intelligent Transportation System
(ITS) in Metro cities for Road Traffic Safety using
VANET Route Traffic Analysis (VRTA),” pp. 383-386,
2016.
[4] B. T. Sharef, R. A. Alsaqour, and M. Ismail, “Vehicular
communication Ad-hoc routing protocols: A survey,”
Journal of Network and Computer Applications, vol. 40,
pp. 363–396, 2014.
[5] S. H. Ahmed, S. H. Bouk, M. A. Yakub, D. Kim, H. Song,
and J. Lloret, “Controlled data and interest evaluation in
vehicular named data networks,” IEEE Transaction on
Vehicular Technology, vol. 65, no. 6, pp. 3954-3963, 2016.
[6] S. S. Joshi and S. R. Birada, “Communication framework
for jointly addressing issues of routing overhead and
energy drainage in MANET,” in Proc. Twelfth
International Multi-Conference on Information Processing,
2016, pp. 57–63.
[7] B. Sahadev, K. S. Zade, and S. H. Sheikh, “Survey on
realistic simulation for comparison of network routing
protocol in VANET,” in Proc. 2nd National Conference on
Innovative Paradigms in Engineering & Technology, 2013,
pp. 26-29.
©2020 Journal of Communications 684
Journal of Communications Vol. 15, No. 9, September 2020
Page 10
[8] M. Ren, L. Khoukhi, H. Labiod, J. Zhang, and V. Veque,
“A mobility-based scheme for dynamic clustering in
vehicular ad-hoc networks (VANETs),” Vehicular
Communications, vol. 9, pp. 233–241, 2017.
[9] X. Hou, Y. Li, M. Chen, D. Wu, D. Jin, and S. Chen,
“Vehicular fog computing: A viewpoint of vehicles as the
infrastructures,” IEEE Transaction on Vehicular
Technology, vol. 65, no. 6, pp. 3860-3873, 2016.
[10] P. Mutalik, S. Nagaraj, J. Vedavyas, R. V. Biradar, and V.
G. C. Patil, “A comparative study on AODV, DSR and
DSDV routing protocols for intelligent transportation
systems in metro cities for road traffic safety using
VANET route traffic analysis (VRTA),” in Proc. IEEE
International Conference on Advances in Electronics,
Communication and Computer Technology, 2016, pp. 383–
386.
[11] M. Patra, R. Thakur, and C. S. Murphy, “Improving delay
and energy efficiency of vehicular networks using mobile
FEMTO access points,” IEEE Transactions on Vehicular
Technology, vol. 66, no. 2, pp. 1496-1505, 2017.
[12] I. Basaran and H. Bulut, “Performance comparison of non-
delay tolerant VANET routing protocols,” in Proc.
International Workshop on Urban Mobility & Intelligent
Transportation Systems, 2016, pp. 1–6.
[13] N. Alsharif and X. Shen, “iCAR-II: Infrastructure-Based
connectivity aware routing in vehicular networks,” IEEE
Transaction on Vehicular Technology, vol. 66, no. 5, pp.
4231-4244, 2017.
[14] B. Hamid and E. E. Mokhtar, “Performance analysis of the
vehicular ad-hoc networks (VANET) routing protocols,
AODV, DSDV and OLSR,” pp. 1–6, 2017.
[15] A. Datta, “Modified Ant-AODV VANET routing protocol
for Vehicular Ad-hoc Network,” IEEE, pp. 1 – 6, 2017.
[16] A. Abuashour and M. Kadock, “Performance improvement
of cluster-based routing protocol in VANET,” IEEE Access,
vol. 5, pp. 15354 – 15371, 2017.
[17] W. Farook, M. Ali-Khan, and S. Rehman, “A multicast
routing protocol for autonomous military vehicles
communication in VANET,” in Proc. 14th International
Burban Conference on Applied Science and Technology,
2017, pp. 699–706.
[18] F. Jameel and M. A. Javed, “On the performance of
cooperative vehicular networks under antenna correlation
at RSU,” International Journal of Electronics and
Communications, vol. 95, pp. 216-225, 2018.
[19] G. Yan and D. B. Rawat, “Vehicles-to-vehicle
connectivity analysis for vehicular ad-hoc networks,” Ad-
Hoc Networks, vol. 58, pp. 25–35, 2017.
[20] S. Khakpour, R. W. Pazzi, and K. EI-Khatib, “Using
clustering for target tracking in vehicular ad-hoc networks,”
Vehicular Communications, vol. 9, pp. 83– 96, 2017.
[21] S. Boussoufa-Lahlah, F. Semchedine, and L. Bouallouche-
Medjkoune, “Geographic routing protocols for vehicular
ad hoc networks (VANETs),” Vehicular Communications,
vol. 11, pp. 20-31, 2018.
[22] A. M. Said, M. Marot, A. W. Ibrahim, and H. Afifi,
“Modeling interactive real-time applications in VANETs
with performance evaluation,” Computer Networks, vol.
104, pp. 66-78, 2016.
[23] J. Shen, C. Wang, A. Wang, X. Sun, S. Moh, and P. Hung,
“Organized topology-based routing protocol in
incompletely predictable ad-hoc networks,” Computer
Communications, vol. 99, pp. 107–118, 2017.
[24] O. S. Oubbati, A. Lakas, F. Zhou, M. Gunes, N. Lagraa,
and M. B. Yagoubi, “Intelligent UAV-assisted routing
protocol for urban VANETs,” Computer Communications,
vol. 107, pp. 93–111, 2017.
[25] V. Jindal and P. Bedi, “An improved hybrid ant particle
optimization (IHAPO) algorithm for reducing travel time
in VANETs,” Applied Soft Computing, vol. 64, pp. 526–
535, 2018.
[26] F. Abbas and P. Fan, “Clustering-based reliable low
latency routing scheme using ACO method for vehicular
networks,” Vehicular Communications, vol. 12, pp. 66-74,
2018.
[27] V. Jindal and P. Bedi, “Reducing waiting time with parallel
preemptive algorithm in VANETs,” Vehicular
Communications, vol. 7, pp. 58 – 65.
[28] R. Skinderowicz, “An improved ant colony system for the
sequential ordering problem,” Computers and Operations
Research, vol. 86, pp. 1–17, 2017.
[29] Y. Meraihi, D. Acheli, and A. Ramdane-cherif, “QoS
performance evaluation of AODV and DSR Routing
Protocols in City VANET Scenarios,” in Proc. 5th
International Conference on Electrical Engineering,
Boumerdes, pp. 1-6, 2017.
[30] S. Chatterjee and S. Das, “Ant colony optimization based
enhanced dynamic source routing algorithm for mobile ad-
hoc network,” Information Sciences, pp. 67 – 90, 2015.
[31] R. Skinderowicz, “An improved ant colony system for the
sequential ordering problem,” Computers and Operations
Research, pp. 1-17, 2017.
[32] A. Louati, S. Elkosantini, S. Darmoul, and L. B. Said, “A
case based reasoning system to control traffic at signalized
intersections,” Procedia Computer Science, vol. 5, pp.
149–154, 2016.
Copyright © 2020 by the authors. This is an open access article
distributed under the Creative Commons Attribution License
(CC BY-NC-ND 4.0), which permits use, distribution and
reproduction in any medium, provided that the article is
properly cited, the use is non-commercial and no modifications
or adaptations are made.
Elias Khoza received his undergraduate
national diploma in Electrical
Engineering from the Tshwane
University of Technology (South Africa)
in 2007 and the BTech degree in
Computer System Engineering from the
Tshwane University of Technology
(South Africa) in 2016; Master of
computing (MCOMP) degree of computing systems engineering
from Tshwane University of Technology (South Africa) in 2020.
He is currently a part-time lecturer at the Tshwane University of
©2020 Journal of Communications 685
Journal of Communications Vol. 15, No. 9, September 2020
Page 11
Technology. His research interests include VANET, big data,
internet of things (IoT), neural networks, and machine learning.
Chunling Tu received the Bachelor
degree of computer science from Tianjin
University of Technology and Education,
China in 2002; MTech and MSc degrees
in Electrical Engineering from Tshwane
University of Technology (South Africa)
and ESIEE Paris University (France) in
2010; DTech and PhD degrees of
Electrical Engineering from the Tshwane
University of Technology and University Paris East, France in
2015. She is currently a senior lecturer at the Tshwane
University of Technology. Her research interests include image
processing, AI, industrial control, machine learning, deep
learning and pattern recognition.
Pius A. Owolawi received his
undergraduate degree in 2001 from the
Federal University of Technology, Akure,
Nigeria, and also bagged his master’s
and PhD Electrical Engineering from the
University of Kwazulu Natal, South
Africa in 2006 and 2010 respectively.
He is currently the Head of the
Department of Computer Systems Engineering, Tshwane
University of Technology, South Africa. His research interests
include RF, Green communication, radiowave propagation
(Microwave/ Millimeter wave systems), Satellite and free space
optical communications, IoT, Embedded systems, Machine
learning and data analytics. Dr. Owolawi was a recipient of
Joint holder of the best paper award for a paper presented at the
2nd international conference on applied and theoretical
information systems research, in Taipei, Taiwan, 2012 and a
recipient of the Vice Chancellor’s teaching Excellence Award,
2015.
©2020 Journal of Communications 686
Journal of Communications Vol. 15, No. 9, September 2020